Hitachi Rail, an innovative railway company globally, is set to upgrade railway maintenance AI and efficiency through NVIDIA deep learning AI. NVIDIA IGX platform is an industrial-grade AI toolkit currently in use across more than 50 countries at Hitachi Rail that will integrate into its HMAX platform. This is expected to benefit significantly from maintenance operations and thereby reduce the time trains spend at stops before the next train starts operating, enhancing the reliability of transit systems.
The five subsystems of the HMAX platform can analyze thousands of sensor and camera inputs in real-time, allowing Hitachi Rail’s clients to identify infrastructure problems, including track maintenance and overhead power line wear and tear, much more quickly. According to the company, real-time analytics estimate that preventive maintenance is seven times cheaper than fixing a failure when it occurs. The AI installed today at Hitachi can decrease train service delays by 20%, reduce maintenance costs by 15%, and lessen energy consumption at depots by 40%.
To this end, Hitachi Rail intends to extend these savings by implementing NVIDIA IGX and NVIDIA Holoscan platforms for quick sensor data processing. As Hitachi Rail CTO and Executive Director Koji Agatsuma says, delayed digital monitoring could be better. At the same time, real-time AI will avoid service disruptions and improve safety while reducing operational costs.
When integrated with NVIDIA’s AI Enterprise software and Hitachi’s HMAX, new AI applications will be developed to enable operators to oversee train fleets and structures efficiently. For instance, Hitachi Rail Communication in the U.K.’s trains send nearly 50,000 pieces of data per fifth of a second so that maintenance alerts and trends can be detected.
Other than trains, cameras operated on Hitachi Rail trains utilize AI and analyze overhead power lines, which generally require ten days to process for recurring problems. With the help of NVIDIA’s technology, data is processed at the edge, transmitting important data to control centers, thus increasing the efficiency and reliability of transit for passengers globally.
For more information, please click here.
